ghisvail / xtensor-python

Python bindings for xtensor

Python bindings for the xtensor C++ multi-dimensional array library.

• xtensor is a C++ library for multi-dimensional arrays enabling numpy-style broadcasting and lazy computing.

• xtensor-python enables inplace use of numpy arrays in C++ with all the benefits from xtensor

  • C++ universal function and broadcasting
  • STL - compliant APIs.
  • A broad coverage of numpy APIs (see the numpy to xtensor cheat sheet).

The Python bindings for xtensor are based on the pybind11 C++ library, which enables seemless interoperability between C++ and Python.

Installation

xtensor-python is a header-only library. We provide a package for the conda package manager.

conda install -c conda-forge xtensor-python

Dependences on xtensor and pybind11

xtensor-python depends on the xtensor and pybind11 libraries

xtensor-python	xtensor	pybind11
master	0.8.1	^2.1.0
0.8.0	0.8.1	^2.1.0
0.7.1	0.7.3	^2.0.0
0.7.0	0.7.2	^2.0.0
0.6.1	0.7.1	^2.0.0
0.6.0	0.7.1	^2.0.0
0.5.0	0.5.0	^2.0.0
0.4.0	0.5.0	^2.0.0
0.3.0	^0.4.1	^2.0.0
0.2.0	^0.2.1	^1.8.1
0.1.0	^0.1.1	^1.8.1

These dependencies are automatically resolved when using the conda package manager.

Usage

xtensor-python offers two container types wrapping numpy arrays inplace to provide an xtensor semantics

• pytensor
• pyarray.

Both containers enable the numpy-style APIs of xtensor (see the numpy to xtensor cheat sheet).

• On the one hand, pyarray has a dynamic number of dimensions. Just like numpy arrays, it can be reshaped with a shape of a different length (and the new shape is reflected on the python side).

• On the other hand pytensor has a compile time number of dimensions, specified with a template parameter. Shapes of pytensor instances are stack allocated, making pytensor a significantly faster expression than pyarray.

Example 1: Use an algorithm of the C++ library on a numpy array inplace.

C++ code

#include <numeric>                        // Standard library import for std::accumulate
#include "pybind11/pybind11.h"            // Pybind11 import to define Python bindings
#include "xtensor/xmath.hpp"              // xtensor import for the C++ universal functions
#include "xtensor-python/pyarray.hpp"     // Numpy bindings

double sum_of_sines(xt::pyarray<double> &m)
{
    auto sines = xt::sin(m);  // sines does not actually hold any value, which are only computed upon access
    return std::accumulate(sines.begin(), sines.end(), 0.0);
}

PYBIND11_PLUGIN(xtensor_python_test)
{
    pybind11::module m("xtensor_python_test", "Test module for xtensor python bindings");

    m.def("sum_of_sines", sum_of_sines, "Computes the sum of the sines of the values of the input array");

    return m.ptr();
}

Python Code

import numpy as np
import xtensor_python_test as xt

a = np.arange(15).reshape(3, 5)
s = xt.sum_of_sines(v)
s

Outputs

1.2853996391883833

Example 2: Create a universal function from a C++ scalar function

C++ code

#include "pybind11/pybind11.h"
#include "xtensor-python/pyvectorize.hpp"
#include <numeric>
#include <cmath>

namespace py = pybind11;

double scalar_func(double i, double j)
{
    return std::sin(i) - std::cos(j);
}

PYBIND11_PLUGIN(xtensor_python_test)
{
    py::module m("xtensor_python_test", "Test module for xtensor python bindings");

    m.def("vectorized_func", xt::pyvectorize(scalar_func), "");

    return m.ptr();
}

Python Code

import numpy as np
import xtensor_python_test as xt

x = np.arange(15).reshape(3, 5)
y = [1, 2, 3, 4, 5]
z = xt.vectorized_func(x, y)
z

Outputs

[[-0.540302,  1.257618,  1.89929 ,  0.794764, -1.040465],
 [-1.499227,  0.136731,  1.646979,  1.643002,  0.128456],
 [-1.084323, -0.583843,  0.45342 ,  1.073811,  0.706945]]

Installation

We provide a package for the conda package manager.

conda install -c conda-forge xtensor-python

This will pull the dependencies to xtensor-python, that is pybind11 and xtensor.

Project cookiecutter

A template for a project making use of xtensor-python is available in the form of a cookie cutter here.

This project is meant to help library authors get started with the xtensor python bindings.

It produces a project following the best practices for the packaging and distribution of Python extensions based on xtensor-python, including a setup.py file and a conda recipe.

Building and Running the Tests

Testing xtensor-python requires pytest

py.test .

To pick up changes in xtensor-python while rebuilding, delete the build/ directory.

Building the HTML Documentation

xtensor-python's documentation is built with three tools

• doxygen
• sphinx
• breathe

While doxygen must be installed separately, you can install breathe by typing

pip install breathe

Breathe can also be installed with conda

conda install -c conda-forge breathe

Finally, build the documentation with

make html

from the docs subdirectory.

License

We use a shared copyright model that enables all contributors to maintain the copyright on their contributions.

This software is licensed under the BSD-3-Clause license. See the LICENSE file for details.